Setup and evaluation of an AI chatbot for Learning Passport customer support

Enable continuous access to education for children, youth, and teachers around the world and drive improved learning outcomes through high-quality, portable education to support their entry into other education or opportunity pathways, including formal education

The Learning Passport platform continues to scale rapidly across multiple country deployments, supporting millions of learners and growing partner ecosystems. With this growth, the volume of first-level customer support tickets has increased significantly, and an analysis of historical tickets shows that a small number of recurring themes (authentication, offline & hub operations, content & media, instance lifecycle) account for the majority of requests. Many of these can be resolved with clear documentation, which already exists but is not consolidated in a single, searchable location. This MVP will test whether an AI chatbot, connected to a well-organized Confluence knowledge base and to the existing JIRA service desk, can deflect a meaningful share of these requests, improve user experience, and free up the technical team to focus on higher-value work.

We are looking for 2 passionate Online Volunteers who, under the guidance of the Programme Specialist (Digital Education), will support our work by setting up and piloting a minimum viable version (MVP) of an AI-powered customer support chatbot for the Learning Passport platform. The chatbot will be configured using the JIRA Service Management Virtual Service Agent, will answer common end-user questions from a curated knowledge base, and will automatically create JIRA tickets for issues that require human intervention. The Online Volunteers will also be responsible for curating and uploading the existing support documentation (currently scattered across multiple repositories) into Confluence as the chatbot's knowledge source, and for evaluating the MVP's effectiveness against a defined set of use cases. In particular, the Online Volunteers will be asked to: - Identify, with the Programme Specialist, 2 to 3 priority use cases to scope the MVP - Review, vet and curate the existing support documentation needed to cover the chosen use cases - Upload and organize the vetted content in Confluence as the chatbot's knowledge base - Configure the JIRA Service Management Virtual Service Agent: intent flows, AI answers from the knowledge base, and ticket routing rules for the chosen use cases - Run structured tests against historical tickets and a sample of live conversations, taking screenshots and recordings as needed - Measure key MVP metrics: deflection rate, accuracy of answers, ticket routing precision, and user satisfaction - Document findings, limitations and recommendations for a possible next phase - Provide regular progress updates and present the final MVP evaluation to the Learning Passport team The Online Volunteers will have the opportunity to build connections with the Learning Passport team at UNICEF, contribute to a concrete digital innovation pilot, and learn about how AI is being applied to scale customer success in education technology.